% TVP-VAR Time varying structural VAR with stochastic volatility
% ------------------------------------------------------------------------------------
% This code implements the TVP-VAR model as in Primiceri (2005). See also
% the monograph, Section 4.2 and Section 3.3.2.
% ************************************************************************************
% The model is:
%
%     Y(t) = B0(t) + B1(t)xY(t-1) + B2(t)xY(t-2) + e(t) 
% 
%  with e(t) ~ N(0,SIGMA(t)), and  L(t)' x SIGMA(t) x L(t) = D(t)*D(t),
%             _                                          _
%            |    1         0        0       ...       0  |
%            |  L21(t)      1        0       ...       0  |
%    L(t) =  |  L31(t)     L32(t)    1       ...       0  |
%            |   ...        ...     ...      ...      ... |
%            |_ LN1(t)      ...     ...    LN(N-1)(t)  1 _|
% 
% 
% and D(t) = diag[exp(0.5 x h1(t)), .... ,exp(0.5 x hn(t))].
%
% The state equations are
%
%            B(t) = B(t-1) + u(t),            u(t) ~ N(0,Q)
%            l(t) = l(t-1) + zeta(t),      zeta(t) ~ N(0,S)
%            h(t) = h(t-1) + eta(t),        eta(t) ~ N(0,W)
%
% where B(t) = [B0(t),B1(t),B2(t)]', l(t)=[L21(t),...,LN(N-1)(t)]' and
% h(t) = [h1(t),...,hn(t)]'.
%
% ************************************************************************************
%   NOTE: 
%      There are references to equations of Primiceri, "Time Varying Structural Vector
%      Autoregressions & Monetary Policy",(2005),Review of Economic Studies 72,821-852
%      for your convenience. The definition of vectors/matrices is also based on this
%      paper.
% ------------------------------------------------------------------------------------

clear all;
clc;
randn('state',sum(100*clock)); %#ok<*RAND>
rand('twister',sum(100*clock));
addpath C:\Users\Annika\Documents\AAMH\Data2022\
%----------------------------------LOAD DATA----------------------------------------
% Load Korobilis (2008) quarterly data
[num] = xlsread('yearlab1980.xlsx');
%Y=num;
yearlab = num;
[num] = xlsread('USA_DataIP.xlsx');
Y=(num);

% Number of observations and dimension of X and Y
t=size(Y,1); % t is the time-series observations of Y
M=size(Y,2); % M is the dimensionality of Y

% Number of factors & lags:
tau = 20; % tau is the size of the training sample
p = 2; % p is number of lags in the VAR part
numa = M*(M-1)/2; % Number of lower triangular elements of A_t (other than 0's and 1's)
% ===================================| VAR EQUATION |==============================
% Generate lagged Y matrix. This will be part of the X matrix
ylag = mlag2(Y,p); % Y is [T x M]. ylag is [T x (Mp)]
% Form RHS matrix X_t = [1 y_t-1 y_t-2 ... y_t-k] for t=1:T
ylag = ylag(p+tau+1:t,:);

K = M + p*(M^2); % K is the number of elements in the state vector
% Create Z_t matrix.
Z = zeros((t-tau-p)*M,K);
for i = 1:t-tau-p
    ztemp = eye(M);
    for j = 1:p        
        xtemp = ylag(i,(j-1)*M+1:j*M);
        xtemp = kron(eye(M),xtemp);
        ztemp = [ztemp xtemp];  %#ok<AGROW>
    end
    Z((i-1)*M+1:i*M,:) = ztemp;
end

% Redefine FAVAR variables y
y = Y(tau+p+1:t,:)';
yearlab = yearlab(tau+p+1:t);
% Time series observations
t=size(y,2);   % t is now 215 - p - tau = 173

%----------------------------PRELIMINARIES---------------------------------
% Set some Gibbs - related preliminaries
nrep = 10000;  % Number of replications �NDRAR H�R
nburn = 30000;   % Number of burn-in-draws
it_print = 100;  %Print in the screen every "it_print"-th iteration

%========= PRIORS:
% To set up training sample prior a-la Primiceri, use the following subroutine
[B_OLS,VB_OLS,A_OLS,sigma_OLS,VA_OLS]= ts_prior(Y,tau,M,p);

% % Or use uninformative values
% A_OLS = zeros(numa,1);
% B_OLS = zeros(K,1);
% VA_OLS = eye(numa);
% VB_OLS = eye(K);
% sigma_OLS = 0*ones(M,1);

% Set some hyperparameters here (see page 831, end of section 4.1)
k_Q = 0.05;
k_S = 0.1;
k_W = 1;

% We need the sizes of some matrices as prior hyperparameters (see page
% 831 again, lines 2-3 and line 6)
sizeW = M; % Size of matrix W
sizeS = 1:M; % Size of matrix S

%-------- Now set prior means and variances (_prmean / _prvar)
% These are the Kalman filter initial conditions for the time-varying
% parameters B(t), A(t) and (log) SIGMA(t). These are the mean VAR
% coefficients, the lower-triangular VAR covariances and the diagonal
% log-volatilities, respectively 
% B_0 ~ N(B_OLS, 4Var(B_OLS))
B_0_prmean = B_OLS;
B_0_prvar = 4*VB_OLS;
% A_0 ~ N(A_OLS, 4Var(A_OLS))
A_0_prmean = A_OLS;
A_0_prvar = 4*VA_OLS;
% log(sigma_0) ~ N(log(sigma_OLS),I_n)
sigma_prmean = sigma_OLS;
sigma_prvar = 4*eye(M);

% Note that for IW distribution I keep the _prmean/_prvar notation....
% Q is the covariance of B(t), S is the covariance of A(t) and W is the
% covariance of (log) SIGMA(t)
% Q ~ IW(k2_Q*size(subsample)*Var(B_OLS),size(subsample))
Q_prmean = ((k_Q)^2)*tau*VB_OLS;
Q_prvar = tau;
% W ~ IG(k2_W*(1+dimension(W))*I_n,(1+dimension(W)))
W_prmean = ((k_W)^2)*ones(M,1);
W_prvar = 2;
% S ~ IW(k2_S*(1+dimension(S)*Var(A_OLS),(1+dimension(S)))
S_prmean = cell(M-1,1);
S_prvar = zeros(M-1,1);
ind = 1;
for ii = 2:M
    % S is block diagonal as in Primiceri (2005)
    S_prmean{ii-1} = ((k_S)^2)*(1 + sizeS(ii-1))*VA_OLS(((ii-1)+(ii-3)*(ii-2)/2):ind,((ii-1)+(ii-3)*(ii-2)/2):ind);
    S_prvar(ii-1) = 1 + sizeS(ii-1);
    ind = ind + ii;
end

%========= INITIALIZE MATRICES:
% Specify covariance matrices for measurement and state equations
consQ = 0.0001;
consS = 0.0001;
consH = 0.01;
consW = 0.0001;
Ht = kron(ones(t,1),consH*eye(M));   % Initialize Htdraw, a draw from the VAR covariance matrix
Htchol = kron(ones(t,1),sqrt(consH)*eye(M)); % Cholesky of Htdraw defined above
Qdraw = consQ*eye(K);   % Initialize Qdraw, a draw from the covariance matrix Q
Sdraw = consS*eye(numa);  % Initialize Sdraw, a draw from the covariance matrix S
Sblockdraw = cell(M-1,1); % ...and then get the blocks of this matrix (see Primiceri)
ijc = 1;
for jj=2:M
    Sblockdraw{jj-1} = Sdraw(((jj-1)+(jj-3)*(jj-2)/2):ijc,((jj-1)+(jj-3)*(jj-2)/2):ijc);
    ijc = ijc + jj;
end
Wdraw = consW*ones(M,1);    % Initialize Wdraw, a draw from the covariance matrix W
Btdraw = zeros(K,t);     % Initialize Btdraw, a draw of the mean VAR coefficients, B(t)
Atdraw = zeros(numa,t);  % Initialize Atdraw, a draw of the non 0 or 1 elements of A(t)
Sigtdraw = zeros(t,M);   % Initialize Sigtdraw, a draw of the log-diagonal of SIGMA(t)
sigt = kron(ones(t,1),0.01*eye(M));   % Matrix of the exponent of Sigtdraws (SIGMA(t))
statedraw = 5*ones(t,M);       % initialize the draw of the indicator variable 
                               % (of 7-component mixture of Normals approximation)
Zs = kron(ones(t,1),eye(M));

% Storage matrices for posteriors and stuff
Bt_postmean = zeros(K,t);    % regression coefficients B(t)
At_postmean = zeros(numa,t); % lower triangular matrix A(t)
Sigt_postmean = zeros(t,M);  % diagonal std matrix SIGMA(t)
Qmean = zeros(K,K);          % covariance matrix Q of B(t)
Smean = zeros(numa,numa);    % covariance matrix S of A(t)
Wmean = zeros(M,1);          % covariance matrix W of SIGMA(t)

sigmean = zeros(t,M);    % mean of the diagonal of the VAR covariance matrix
cormean = zeros(t,numa); % mean of the off-diagonal elements of the VAR cov matrix
sig2mo = zeros(t,M);     % squares of the diagonal of the VAR covariance matrix
cor2mo = zeros(t,numa);  % squares of the off-diagonal elements of the VAR cov matrix

%========= IMPULSE RESPONSES:
% Note that impulse response and related stuff involves a lot of storage
% and, hence, put istore=0 if you do not want them
istore = 1;
if istore == 1;
    nhor = 21;  % Impulse response horizon
    imp75 = zeros(nrep,M,nhor);
    imp81 = zeros(nrep,M,nhor);
    imp96 = zeros(nrep,M,nhor);
    bigj = zeros(M,M*p);
    bigj(1:M,1:M) = eye(M);
end
%----------------------------- END OF PRELIMINARIES ---------------------------

%====================================== START SAMPLING ========================================
%==============================================================================================
tic; % This is just a timer
disp('Number of iterations');

for irep = 1:nrep + nburn    % GIBBS iterations starts here
    % Print iterations
    if mod(irep,it_print) == 0
        disp(irep);toc;
    end
    % -----------------------------------------------------------------------------------------
    %   STEP I: Sample B from p(B|y,A,Sigma,V) (Drawing coefficient states, pp. 844-845)
    % -----------------------------------------------------------------------------------------


    draw_beta
    
    %-------------------------------------------------------------------------------------------
    %   STEP II: Draw A(t) from p(At|y,B,Sigma,V) (Drawing coefficient states, p. 845)
    %-------------------------------------------------------------------------------------------
    
    
    draw_alpha
    
    
    %------------------------------------------------------------------------------------------
    %   STEP III: Draw diagonal VAR covariance matrix log-SIGMA(t)
    %------------------------------------------------------------------------------------------
    
    draw_sigma
    
    
    % Create the VAR covariance matrix H(t). It holds that:
    %           A(t) x H(t) x A(t)' = SIGMA(t) x SIGMA(t) '
    Ht = zeros(M*t,M);
    Htsd = zeros(M*t,M);
    for i = 1:t
        inva = inv(capAt((i-1)*M+1:i*M,:));
        stem = diag(sigt(i,:));
        Hsd = inva*stem;
        Hdraw = Hsd*Hsd';
        Ht((i-1)*M+1:i*M,:) = Hdraw;  % H(t)
        Htsd((i-1)*M+1:i*M,:) = Hsd;  % Cholesky of H(t)
    end
    
    %----------------------------SAVE AFTER-BURN-IN DRAWS AND IMPULSE RESPONSES -----------------
    if irep > nburn;               
        % Save only the means of parameters. Not memory efficient to
        % store all draws (at least for the time-varying parameters vectors,
        % which are large). If you want to store all draws, it is better to
        % save them in a file at each iteration. Use the MATLAB command 'save'
        % (type 'help save' in the command window for more info)
        Bt_postmean = Bt_postmean + Btdraw;   % regression coefficients B(t)
        At_postmean = At_postmean + Atdraw;   % lower triangular matrix A(t)
        Sigt_postmean = Sigt_postmean + Sigtdraw;  % diagonal std matrix SIGMA(t)
        Qmean = Qmean + Qdraw;     % covariance matrix Q of B(t)
        ikc = 1;
        for kk = 2:M
            Sdraw(((kk-1)+(kk-3)*(kk-2)/2):ikc,((kk-1)+(kk-3)*(kk-2)/2):ikc)=Sblockdraw{kk-1};
            ikc = ikc + kk;
        end
        Smean = Smean + Sdraw;    % covariance matrix S of A(t)
        Wmean = Wmean + Wdraw;    % covariance matrix W of SIGMA(t)
        % Get time-varying correlations and variances
        stemp6 = zeros(M,1);
        stemp5 = [];
        stemp7 = [];
        for i = 1:t
            stemp8 = corrvc(Ht((i-1)*M+1:i*M,:));
            stemp7a = [];
            ic = 1;
            for j = 1:M
                if j>1;
                    stemp7a = [stemp7a ; stemp8(j,1:ic)']; %#ok<AGROW>
                    ic = ic+1;
                end
                stemp6(j,1) = sqrt(Ht((i-1)*M+j,j));
            end
            stemp5 = [stemp5 ; stemp6']; %#ok<AGROW>
            stemp7 = [stemp7 ; stemp7a']; %#ok<AGROW>
        end
        sigmean = sigmean + stemp5; % diagonal of the VAR covariance matrix
        cormean =cormean + stemp7;  % off-diagonal elements of the VAR cov matrix
        sig2mo = sig2mo + stemp5.^2;
        cor2mo = cor2mo + stemp7.^2;
         
        if istore==1;
            
            
            IRA_tvp
            
            
        end %END the impulse response calculation section   
    end % END saving after burn-in results 
end %END main Gibbs loop (for irep = 1:nrep+nburn)
clc;
toc; % Stop timer and print total time
%=============================GIBBS SAMPLER ENDS HERE==================================
Bt_postmean = Bt_postmean./nrep;  % Posterior mean of B(t) (VAR regression coeff.)
At_postmean = At_postmean./nrep;  % Posterior mean of A(t) (VAR covariances)
Sigt_postmean = Sigt_postmean./nrep;  % Posterior mean of SIGMA(t) (VAR variances)
Qmean = Qmean./nrep;   % Posterior mean of Q (covariance of B(t))
Smean = Smean./nrep;   % Posterior mean of S (covariance of A(t))
Wmean = Wmean./nrep;   % Posterior mean of W (covariance of SIGMA(t))

sigmean = sigmean./nrep;
cormean = cormean./nrep;
sig2mo = sig2mo./nrep;
cor2mo = cor2mo./nrep;

% Standard deviations of residuals of Inflation, Unemployment and Interest Rates
figure
set(0,'DefaultAxesColorOrder',[0 0 0],...      
    'DefaultAxesLineStyleOrder','-|-|-')
subplot(3,1,1)
plot(yearlab,sigmean(:,1))
title('Posterior mean of the standard deviation of residuals in Inflation equation')
subplot(3,1,2)
plot(yearlab,sigmean(:,2))
title('Posterior mean of the standard deviation of residuals in Unemployment equation')
subplot(3,1,3)
plot(yearlab,sigmean(:,3))
title('Posterior mean of the standard deviation of residuals in Interest Rate equation')

if istore == 1 
    qus = [.16, .5, .84];
    imp75XY=squeeze(quantile(imp75,qus));
    imp81XY=squeeze(quantile(imp81,qus));
    imp96XY=squeeze(quantile(imp96,qus));
            
    % Plot impulse responses
    figure       
    set(0,'DefaultAxesColorOrder',[0 0 0],...
        'DefaultAxesLineStyleOrder','--|-|--')
    subplot(3,3,1)
    plot(1:nhor,squeeze(imp75XY(:,1,:)))
    title('Impulse response of inflation, 1975:Q1')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)
    subplot(3,3,2)
    plot(1:nhor,squeeze(imp75XY(:,2,:)))
    title('Impulse response of unemployment, 1975:Q1')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)    
    subplot(3,3,3)
    plot(1:nhor,squeeze(imp75XY(:,3,:)))
    title('Impulse response of interest rate, 1975:Q1')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)    
    subplot(3,3,4)
    plot(1:nhor,squeeze(imp81XY(:,1,:)))
    title('Impulse response of inflation, 1981:Q3')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)    
    subplot(3,3,5)
    plot(1:nhor,squeeze(imp81XY(:,2,:)))
    title('Impulse response of unemployment, 1981:Q3')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)    
    subplot(3,3,6)
    plot(1:nhor,squeeze(imp81XY(:,3,:)))
    title('Impulse response of interest rate, 1981:Q3')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)    
    subplot(3,3,7)
    plot(1:nhor,squeeze(imp96XY(:,1,:)))
    title('Impulse response of inflation, 1996:Q1')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)    
    subplot(3,3,8)
    plot(1:nhor,squeeze(imp96XY(:,2,:)))
    title('Impulse response of unemployment, 1996:Q1')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)
    subplot(3,3,9)
    plot(1:nhor,squeeze(imp96XY(:,3,:)))
    title('Impulse response of interest rate, 1996:Q1')
    xlim([1 nhor])
    set(gca,'XTick',0:3:nhor)
end
    qus = [.05, .5, .95];

% f�rst IP:s irf map XR
    
    PT1=quantile(irf_IP,qus,2);
    PT2=squeeze(PT1);
    PT_Usa=PT2(:,:,4);
    %    Koll = squeeze(Koll1);
  % plot(yearlab,Koll);
  %plot(Koll)
  Dates=[1;yearlab];

col_header = {'Dates','PT_l','PT','PT_u'} 


%Usa_data = [Pstar_all(:,1),CPI_all(:,1),XR_all(:,1)];
filename = 'Usa_PT_IP.xlsx';
writematrix(Dates,filename)
xlswrite('Usa_PT_IP.xlsx',PT_Usa,'B2:D147');
xlswrite('Usa_PT_IP.xlsx',col_header,'A1:D1');

% sedan vanlig PT; irf_pi/irf_xr

PT_standard = irf_IP./irf_XR;
    
    PT1=quantile(PT_standard,qus,2);
    PT2=squeeze(PT1);
    PT_Usa_standard=PT2(:,:,4);
    %    Koll = squeeze(Koll1);
  % plot(yearlab,Koll);
  %plot(Koll)
  Dates=[1;yearlab];

col_header = {'Dates','PT_standard_l','PT_standard','PT_standard_u'} 


%Usa_data = [Pstar_all(:,1),CPI_all(:,1),XR_all(:,1)];
filename = 'Usa_PT_standard.xlsx';
writematrix(Dates,filename)
xlswrite('Usa_PT_standard.xlsx',PT_Usa_standard,'B2:D147');
xlswrite('Usa_PT_standard.xlsx',col_header,'A1:D1');


% sedan xr:s AR

  qxr = [.5];
    
    XR1=quantile(irf_XR,qxr,2);
    XR2=squeeze(XR1);
    no_obs = size(yearlab,1);

    halflifeXR = zeros(no_obs,1);
    for k=1:no_obs
    XR3=XR2(k,:);% �r irf f�r datum k

  % hitta det intervall d�r halflife finns
  Index1 = find(XR3<0.5,1);

  interval = linspace(XR3(Index1-1),XR3(Index1),1000);
  % ger 100 punkter
  Index2 = find(interval<0.5,1);      
  halflifeXR(k)=Index1-1+Index2*0.001;   
   
    end % k, obs
 

  Dates=[1;yearlab];

col_header = {'Dates','AR_xr'} 


%USA_data = [Pstar_all(:,1),CPI_all(:,1),XR_all(:,1)];
filename = 'USA_AR_xr.xlsx';
writematrix(Dates,filename)
xlswrite('USA_AR_xr.xlsx',halflifeXR,'B2:B147');
xlswrite('USA_AR_xr.xlsx',col_header,'A1:B1');


% sist IP:s ar

  qxr = [.5];
    
    AR1=quantile(AR_IP,qxr,2);
    AR2=squeeze(AR1);


    halflifeIP = zeros(no_obs,1);
    for k=1:no_obs
    AR3=AR2(k,:);% �r irf f�r datum k

  % hitta det intervall d�r halflife finns
  Index1 = find(AR3<0.5,1);

  interval = linspace(AR3(Index1-1),AR3(Index1),1000);
  % ger 100 punkter
  Index2 = find(interval<0.5,1);      
  halflifeIP(k)=Index1-1+Index2*0.001;   
   
    end % k, obs
 


  Dates=[1;yearlab];

col_header = {'Dates','AR_IP'} 


%USA_data = [Pstar_all(:,1),CPI_all(:,1),XR_all(:,1)];
filename = 'USA_AR_IP.xlsx';
writematrix(Dates,filename)
xlswrite('USA_AR_IP.xlsx', halflifeIP,'B2:B147');
xlswrite('USA_AR_IP.xlsx',col_header,'A1:B1');


col_header = {'Dates','sig_IP', 'sig_xr'} 
% nu sigma ip och xr
sig_out = sigmean(:,2:3);
filename = 'Usa_sigma_IP_XR.xlsx';
writematrix(Dates,filename)
xlswrite('Usa_sigma_IP_XR.xlsx',sig_out,'B2:C146');
xlswrite('Usa_sigma_IP_XR.xlsx',col_header,'A1:C1');

disp('             ')
disp('To plot impulse responses, use:')
disp('plot(1:nhor,squeeze(imp75XY(:,VAR,:))), for impulse responses at 1975:Q1')
disp('plot(1:nhor,squeeze(imp81XY(:,VAR,:))), for impulse responses at 1981:Q3')
disp('plot(1:nhor,squeeze(imp96XY(:,VAR,:))), for impulse responses at 1996:Q1')
disp('             ')
disp('where VAR=1 for impulses of inflation, VAR=2 for unemployment and VAR=3 for interest rate')



